Linear model for 3D motion estimation and shape reconstruction based on the straight-line optical flow

Abstract

In this paper, a novel linear three-dimensional(3D) motion estimation and shape reconstruction method is proposed based on the straight-line optical flow. Firstly, the expression of the straight-line and the definition of the straight-line optical flow are presented. Then, a novel linear 3D motion estimation model is presented based on the straight-line optical flow, which just requests at least two corresponding straight-lines between continues two frames to compute the rotation and translation parameters of the moving object. With the estimated motion parameters, the 3D shape of the object could be reconstructed effortlessly. Finally, the elaborate experiments with the synthetic and real image sequences show the proposed method has the good performance for motion estimation and shape reconstruction.